Multi-level Distributed Arithmetic Coding with Nested Lattice Quantization

In this paper, a scheme based on 1-D nested lattice quantization followed by multi-level distributed arithmetic coding (MLDAC) as the Slepian-Wolf (SW) code is proposed for the lossy source coding of continuous sources. This system can be employed in distributed video and image coding applications. The output of the quantizer is first converted to binary, and then the SW coding is applied on each bit plane. An efficient algorithm for joint decoding of bit planes at the decoder is proposed, which exploits the dependency of the bit planes with each other and the side-information available at the decoder. Also, a methodology for rate allocation among the bit planes is presented. The quantization parameters are designed to improve the end-to-end system performance based on its overall rate-distortion function taking into account both the distortions due to quantization and imperfect practical Slepian-Wolf coding. The simulation results demonstrate the effectiveness of the proposed MLDAC with the lattice quantization.